The 9 Discovery Questions That Actually Predict Deal Outcomes (And the 40 That Don't)
A ranked breakdown of which discovery questions actually correlate with progressed and won deals — and why most 'best questions' lists are recycled folklore.
Search "best discovery questions for sales calls" and you'll get the same forty questions in a different running order, recycled across a decade of sales blogs, none of them ever tested against an actual deal outcome. Nobody pulled the transcripts. Nobody checked whether the calls that opened with "what keeps you up at night" closed at a different rate than the ones that didn't. It's folklore wearing a framework's clothes, and reps memorise it because a laminated list feels like preparation.
I've spent fifteen years running discovery calls and grading other people's, and for the last couple of years we've been doing it at scale — thousands of assessed sales calls, tagged against outcome, run through The Mastery Standard's discovery competency. That's a different exercise from "here are nine questions I like." It lets you ask which questions actually correlate with progressed and won deals, and which ones just sound good in a training deck.
The pattern that falls out is simple and slightly uncomfortable. A question predicts outcome when it costs the buyer something to answer honestly — a number, a name, a date, an admission that something failed before, something a colleague or a competitor could check them on later. A question predicts nothing when the buyer can answer it for free, with a generic or flattering non-answer that commits them to nothing at all. Most of the popular list is the second kind.
The 9 that hold up
| Rank | Question pattern | What it actually tests | Why it predicts |
|---|---|---|---|
| 1 | "What happens if you don't fix this by [date]?" | Cost of inaction, in the buyer's own words | Forces a real consequence or a real number onto the record; a vague answer here usually means there's no real deadline at all |
| 2 | "Who else has to sign off, and what's the one thing they'll ask that you can't answer yet?" | Coalition and paper process | Surfaces the stakeholder gap while it's still cheap to close, instead of at contract review |
| 3 | "Walk me through the last time you tried to fix this and it stalled — what actually killed it?" | Organisational precedent | Most no-decision losses are repeats of a specific prior failure; this finds out if you're about to repeat one |
| 4 | "What number does this need to hit to be worth doing over doing nothing?" | Economic threshold | Separates a wish list from a funded problem — a shrug here means there's no real budget yet |
| 5 | "If you took this to your boss with nothing else, what's the first hole they'd poke in it?" | Decision-criteria depth | Tests whether the buyer has pressure-tested their own case, or is just repeating what you told them |
| 6 | "Whose budget does this come out of, and have they said yes, or just that it sounds interesting?" | Budget ownership vs. budget existence | "Interesting" and "funded" get conflated constantly; this question splits them apart |
| 7 | "What did you like about how [competitor] ran their process, compared to us?" | Competitive frame, said out loud | Buyers rarely volunteer this unprompted; asked directly, it's the most under-used question on the list |
| 8 | "If this doesn't happen, what do you do instead?" | Status quo cost | A credible alternative means real competition for the deal; a shrug means it's more winnable than the pipeline notes suggest |
| 9 | "Tell me about the last purchase like this that got killed after everyone agreed it made sense." | Process risk | Names the ghost in the room before it kills your deal the same way |
Notice what's absent from that table: nothing about rapport, nothing hypothetical, nothing the buyer can answer with a straight face and no cost. Every one of these makes the buyer expose something — a failure, a weakness in their own case, a name, a number — that a lazy answer would visibly dodge.
Why the other 40 are theatre
The rejected pile sorts into a handful of repeat offenders:
- Rapport openers dressed as discovery — "tell me about your role," "how long have you been at the company." Fine for warming up a call. Zero correlation with outcome, because there's no wrong answer and no cost to a generic one.
- Hypotheticals — "if you had a magic wand, what would you fix?" Buyers enjoy these because they're free fantasy, disconnected from budget, timeline, or actual authority to act.
- Self-report scales — "on a scale of one to ten, how important is solving this?" Everyone says seven or eight. It's a number with no information in it, because there's no consequence to inflating it.
- "What keeps you up at night" — the single most recycled line in sales training, and one of the least diagnostic. Every buyer has heard it a hundred times and has a rehearsed answer ready.
- Timeline taken at face value — "what's your timeline for this?" with no follow-up. A date offered with no pressure behind it is decoration, not information.
None of these are bad questions to ask a friend. They're just questions a buyer can answer without revealing anything you can act on, which is exactly why they don't shift the number on outcome.
What this means for how you run discovery
This isn't an argument for a longer list — it's an argument for a shorter one, asked properly, with the discipline to sit in the silence that follows a question that actually costs the buyer something. Most reps ask a good question and then rescue the buyer from the pause with an easier second question before the first one has landed. The skill was never the question bank. It's asking one of the nine, then shutting up.
If you want to check your own calls against this before your manager does, the Discovery Questions Cheat Sheet maps each of the nine patterns to phrasing you can adapt without sounding like you're reading a script, and the Qualification IQ Self-Assessment Quiz is a fast way to see where your own instinct diverges from what actually predicts outcome. Managers running call reviews might get more out of the Discovery Call Scorecard for Managers than another adjective-heavy competency model — it scores for the mechanism above, not for politeness.
Ask fewer questions. Ask ones that cost something to answer honestly. Then wait.